KICKING THE HABIT: The exploitation of collocation in literature and humour. Alan Partington University of Camerino (MC) Italy The power of collocation usually passes unnoticed. It is the invisible glue of text. In many ways, therefore, the effects of normal collocation are best seen when, for some reason, it is missing. There are a number of occasions when the rules - or regularities, or habits - of normal collocation are upset. In some types of literature, for example, authors attempt to investigate and exploit the framework of habit that collocation imposes on language. Similarly, a principal source of humour is the play upon collocation with the aim of upsetting the linguistic expectations of an audience. In this piece of work, by looking at fragments of texts from the areas of literature and humour, I want to see whether anything interesting can be discovered about the general processes of collocation when language users are "bending the rules". Definitions of collocation. The term "collocation" is used in linguistics in two distinct ways. The first of these we might call the "statistical" definition, which lays stress on the frequency with which a particular word, the "node" word, is found in the company of other words, or collocates, over large bodies of text. This is the meaning referred to by Hoey when he states: "Collocation has long been the name given to the relationship a lexical item has with items that appear with greater than random probability in its (textual) context" (1991: 6-8) Leech, in his discussion of "Seven Types of Meaning", talks of "collocative meaning", which he defines as follows: "Collocative meaning consists of the associations a word acquires on account of the meanings of words which tend to occur in its environment." (1978, 20) These statements contain a good working definition of the concept for those studying in the field of corpus linguistics, where large quantities of text can be put on computer and made available for analysis. The habitual associations of a word with other items can thus be studied both by calling up concordances of that word and by obtaining lists of the most frequent collocates of the node.